Author's response to self-directed learning letters.

To the Editor: I thank Drs Guglielmino, Bulik, and Shokar for their reactions 1,2 to my commentary 3 on self-directed learning (SDL). In spite of what seems to have been the perception of these authors, however, my initial commentary was not so much a critique of the Self-directed Learning Readiness Scale (SDLRS) as a strong encouragement to explore avenues of research beyond it. Since the SDLRS has become the focus of the debate, it may be appropriate for me to relate my own experience with it and to explain why I came to the conclusion that other means of quantifying SDL preparedness are direly needed. I tried on three different occasions to use the SDLRS and derived tests (like the Learning Preference Assessment test [LPA]) with graduate students. Every time, I met with strong opposition. Part of the stu-dents' reaction stemmed from concern about what I would do with the test scores. Even after I reassured them in this respect, however, each group of students endeavored to take the test apart and pointed out how potentially misleading each question could be. Statistical arguments I presented, similar to those mentioned by Drs Guglie lmino, Bulik, and Shokar, were quickly dismissed on the grounds that it would probably be relatively easy to find some other statistical test (eg, Bayesian or nonparametric), different than any used to date in connection with the SDLRS but equally valid theoretically, which would give a very different outcome concerning the significance of 20-or even 30-points differences in test scores among various populations. In that, the students echoed some of the current and frequently heated debate on significance testing in the scientific community. Their opinion was borne out by my own observations of their SDL readiness. Even though the students scored high on the LPA, well above the average of the general adult population, they turned out to be generally unprepared for the kind of SDL that they will be expected to engage later in their career. Even as a " snapshot " of the students' SDL readiness, I found the LPA scores misleading and of little value. These students also pointed out that, leaving statistical arguments aside, one might consider that for environmental scientists who will need to constantly update their knowledge base over their whole career, a 10% score difference on a test that involves questions like " I love to learn " or " I …